Bit funny how the article states that the systems are unobtrusive as to be essentially useless, which ignores the fact that these systems are defanged to appease entitled speeders.
Governments around the world impose various requirements on vehicle speeds, most notably ebikes being capped at 20mph for "safety reasons". Meanwhile private cars can travel at over 100mph for no apparent reason other than to not upset drivers at the expense of everyone else on the road.
I don't see the point in arguing in bad faith. Bikes are allowed on cycle paths or even the same path as pedestrians for long stretches. It's not the same. Obviously.
Mopeds can only operate on streets, they cant go in bikelanes or sidewalks. They are subject to the same restrictions as cars (licensure, registration) except that they uniquely have a speed limit
Not having a speed limit is unique only to private car ownership
No, you've got it backwards. NYC categorises mopeds by top speed, where class A don't have a speed limit.
As a secondary issue, you're claiming an NYC rule (that you've misconstrued) is evidence of pro-car rules worldwide, and also you're omitting motorbikes (and class A mopeds) even in NYC.
The NYC classification allow relaxations of rules for class B and C mopeds. No helmet required for class C. No inspections required for B or C. These are pro-moped rules. It seem really unobjective to see them as evidence of pro-car thinking.
The limit in Germany is 25kmh, which translates to around 15.5 mph. And yet we have quite some unlimited Autobahns.
The 10% fastest cars on the Autobahn section near my home go 200kmh/125mph on average. You can guess what the top 5% or 1% drive.
It looks really interesting, but as an experienced climber I'm not sure if just watching a video of my avatar climbing would really help with skill acquisition.
Also, this claims that the wall type or video quality doesn't matter, but I have a hard time understanding how the model would be able to understand that a small crimp could possibly be dual textured and therefore has only a few specific ways of approaching it.
So it seems that this is more for visualizing a climb (which is a skill most climbers should develop) and not really for dialing in some sort of microbeta for a problem.
I suspect that, absent that information about the exact right way to grab a hold, or the exact way to put a foot on a hold, you'll be limited to beta suggestions, which is fine, I think. It'd be like having a group of other climbers nearby to suggest different beta, even if you don't have any friends.
So, in terms of solving complicated beta faster, I see real utility to this.
It is very interesting that since the AI climber is trained on actual climbers, it could, in principle, provide beta to climb consistent with your own style. If you train the bot exclusively on footage of yourself, it would return movement based on your style. If your style is finessy-all-backstep-all-the-time (aka The Edlinger), it can provide beta consistent with that. If your style is to square up and pull (otherwise known as The American), it can provide beta consistent with that instead.
> It is very interesting that since the AI climber is trained on actual climbers, it could, in principle, provide beta to climb consistent with your own style. If you train the bot exclusively on footage of yourself, it would return movement based on your style. If your style is finessy-all-backstep-all-the-time (aka The Edlinger), it can provide beta consistent with that. If your style is to square up and pull (otherwise known as The American), it can provide beta consistent with that instead.
I would think this is actually a Bad Thing. It's very easy to get stuck trying to make a sequence fit your style of climbing. The better approach (especially for long term skill acquisition) is a willingness to learn new styles. That's to say that every sequence is only solvable via one particular style, but I think long term development is hindered if you approach every crux with the one thing you are good at.
> So, in terms of solving complicated beta faster, I see real utility to this.
I can agree with this. But, to the point that others have made, I do wonder what this and the availability of beta videos for many, many routes and blocs does to climbing skill overall. Perhaps I'm just a grumpy old man, but, particularly when bouldering, sorting out the beta should be part of the journey toward eventually sending. Last fall, I visited Hueco Tanks after a six year absence. I suppose I was a bit disappointed to see so many people watching YouTube beta videos of nearly every problem they tried.
>I can agree with this. But, to the point that others have made, I do wonder what this and the availability of beta videos for many, many routes and blocs does to climbing skill overall.
That's a fair concern. That said, there are certain sequences that I'm physically incapable of doing without a dedicated multi-month program of stretching beforehand. Turns out falling on a thumbs-down jam is close enough to a shoulder dislocation that I maybe should've done some PT about it. The extreme is, obviously, the Edlinger vs. American example, but I think the middle ground actually addresses people's peculiar body geometry and/or range of motion. Alex Honnold's exact hip or elbow position might not be as meaningful for Ashima, even if they're on the exact same route.
Such nuances have made a difference in some cases (the one that springs to mind is Todd Skinner's observation that Steve Petro's hips sagged just a little mid-crux on Fiddler on the Roof, which, until corrected, had prevented Petro from nabbing the first ascent of one of the hardest cracks ever climbed to that point). Probably a net-bad for folks projecting Midnight Lightning or similar, but definitely useful for somebody looking to repeat Silence or whatever.
Agreed - so much about the detail of how you would climb something comes down to details that would be hard to measure with a camera, like textures, your estimate of friction, etc. Very cool idea though, looks fun to test.
And this is deep into "sport climbing", borderline gym rat territory imho. It doesn't model all the other core aspects such as protection, rope management, exposure and rest stops. I imaging if you pointed this at a real cliff and recorded several assents it would quickly become a blurry-twitchy mess as all the movements not touching the rock spoiled the data. Maybe for bouldering, but not for real climbing.
There are lots more variables to consider, particularly in lead climbing, even when you have a bolted route. And trad climbing is even more complex than that.
This one-upmanship is very much part of climbing culture. From the top down, the hierarchy is: free soloists (climbing at height without any protection), trad climbers (carry the rope up with them, set their own protection), sports lead climbers (carry the rope up with them, bolts are set into the wall), top rope climbers (rope already there dangling from the top). Right down the bottom, you have aid climbing, when you use equipment to haul yourself up on a rope.
And then there's indoors vs. outdoors, with some dedicated outdoor climbers regarding anything done in a climbing gym as "not real climbing".
Most people don't take this literally, and it's generally considered to be a standing joke in climbing. Sadly, though, some people take it very seriously.
Aid climbing isn't respected on smaller or well-established cliffs, but most understand that it has a place. All the great routes started as aid routes, only being climbed "free" years later. Aid is also a very useful skill in the rain or rescue situations when friction disappears.
Such differences are actually at the heart of climbing. Is taking a helicopter up to the top climbing? Of course not. So from day one it is not about getting to the top but about following invented rules governing how you get to the top. How about a bolted-on ladder? Or pre-placed protection (bolts)? Clean trad, leaving the rock as you found it, is generally seen as the highest form.
(Climbing rock without ropes may be more "pure" but is so dangerous that it should never be idolized.)
What could be interesting is if you could compare your attempts to the avatar climbing and receiving feedback afterwards. This would effectively be a step up from simply recording your send attempts.
I don't see the shape of the holds being a big problem. With some help from indoor companies and hold makers, figuring out which hold model is on the wall should be possible.
As for the usefullness of the software, I'm sceptical too as it don't really solve a problem. But maybe I'm not seeing it and it could be good for beginners :) A good improvement would be adding a comparison between you and the model in term of body position and fluidity of movements.
The idea of incorporating actual hold data and "recognizing" specific holds is interesting, but I'm not sure it completely solves the problem.
The "Boss" from Pusher is arguably the most famous climbing hold ever made. For a decade or more, every gym had one, but they were all unique. Lots of them had micro chips that became critical to usage of the hold. Some had decent texture and some were glassy smooth from years and years and years of use. A lot of the accidental variation in new holds has gone away as the industry has standardized around a handful of industrial fabricators like Aragon, but even over the course of a single indoor boulder problem's life, the accumulation of chalk, sweat, and shoe rubber can have a significant impact on how a hold climbs.
I guess the real question is, do these changes just make routes harder or do they make them fundamentally different? Do they actually change the set of moves that constitutes the easiest way to the top? To be honest, I'm not entirely sure. But it's something interesting to think about.
Exactly, holds will evolve as they get used and more polished, even indoors. Climbing a Moonboard with a new set of holds is quite different than climbing on one with older more polished holds, even if it's the exact same problem and the same holds.
It's an interesting project and it could be fun to watch, but it's completely useless.
To me, the customer here would be climbing gyms, offering a service to climbers.
1. Set up camera on routes
2. Record all climbs
3. Reason through hold details
4. Generate potential movements
5. Show climbs vs ghost movements
6. Feedback to tune model
3 being accomplished by reasoning "if a movement should be possible using the identified hold, but no one successfully does it, the hold must be misidentified or have different properties."
The point in this thread seemed to be "real world holds have different properties, and that defines possible approaches to holds."
To which I pointed out that, with enough data, you could reason backwards to figure out their properties.
Assuming that's solved, if the question is "What is the point?" then I'd answer the same point as golf swing analysis -- structured comparison feedback for continual improvement.
"Have you thought about trying X move at Y point?" or "You're trying X move at Y point, but here's how you differ from someone successfully doing it" both seem useful feedback.
And essentially what's manually generated now, from someone watching and then providing feedback.
With regards to strength, hell, if you wanted to get fancy you could also deduce a specific user's strength, comparing their moves against others' moves on the same features.
> I don't see the shape of the holds being a big problem. With some help from indoor companies and hold makers, figuring out which hold model is on the wall should be possible.
Even if you know the exact hold model and it’s in pristine condition, it’s basically impossible to tell how it’s gonna work from a single angle on video at a distance. Even tiny variations in angle of the wall and rotation of the hold on the wall can completely change how you use it.
Of the full distribution of possible video qualities one can take on a modern phone camera, the vast majority of video qualities will be fine for the AI to understand fine details. Obviously, if you somehow or for some reason, take a video with really bad quality, it will not give you what you want.
Same explanation goes for the walls. If you take a video of just a really dark wall with really bad holds, it is probably won't give you what you want either.
NYPD sends undercover cops across state lines to spy on other states' residents that are engaged in lawful activities. It's well documented.
The NYC government has traditionally treated their police force as its own personal millitary and intelligence service, and neighboring states as hostile foreign enemies.
Comparing YC to Bell Labs, which had deep ties to academia, misses the entire philosophy of YC and Paul Graham. He constantly talks about academia being a hive of orthodoxy, and positions "smart people" as those who can think around it.
By his logic, Bell Labs would be one of the largest hives of orthodox thinking to ever exist, and its the complete antithesis of vc heterodox thought
Um, I don’t think I was advocating any particular level of moderation, was I? More like visibility into its processes amd motivations, and that providing those convincingly and to an appropriate extent is the moderators’ responsibility. Dang’s please-stop-this essays here come to mind, for example—even if not everybody can be dang and not every community would be moved by such essays.
(There are moderation practices that I disapprove of and are not coincidentally outright incompatible with the view I expressed. Like the advice to just ban the user if you dislike interacting with them or if they’re complaining about suppression—especially in a small community like that advice was targeted at, I know I’d be more or less unsalvageably bitter after witnessing this in practice, let alone being its target. But it’s still not the strength that upsets me in this hypothetical, it’s more the perceived arbitrariness. Which, if the moderator is not in fact being arbitrary, is again a communication problem, not a policy one.)
Moderation is overhead, but so is Postgres. Both are very useful solutions to real and difficult problems. Both still have to pay for themselves with some mix of user-visible shinies and keeping out of the way instead of grumbling about how difficult the problem is. The correct choice of that mix is highly situational and I don’t pretend to have the panacea in that respect.
> Um, I don’t think I was advocating any particular level of moderation, was I? More like visibility into its processes amd motivations, and that providing those convincingly and to an appropriate extent is the moderators’ responsibility.
If you are not advocating for a particular type of moderation then why are you all bothered about how any type of moderation is applied? What would be the point of your suggestion?
It's not a "Tangential remark" if your tangent is 10 times longer than your commentary on the article!
> Please don't complain about tangential annoyances—e.g. article or website formats, name collisions, or back-button breakage. They're too common to be interesting.
> Supportive. We build models to support our users, not replace them. We are focused on efficient, specialized, and practical AI performance – not a quest for god-like intelligence. We develop tools that help everyday people and everyday firms use AI to unlock creativity, boost their productivity, and open up new economic opportunities.
Refreshing take on the peak alarmism we see from tech "thought leaders"
This is just marketing. They're positioning themselves as somehow "more human" while building the exact same technology. When a model supports me by doing the work I'd otherwise hire someone to do, the model just replaced someone. And this goes without saying, but a large amount of outsourced tasks today don't exactly require "god-like intelligence".
That was probably said about the automobile, when it replaced horses, or about electrical lamps, when replaced oil-based lamps, no?
I mean, every city had an army of people to light up and down oil lamps in the streets, and these jobs went away. But people were freed up to do better stuff.
It is different this time. I bet that was also said when the transformations that you mentioned occurred, but this time it really is different.
LLM models are pretty general in their capabilities, so it is not like the relatively slow process of electrification, when lamplighters lost their jobs. Everyone can lose their jobs in a matter of months because AI can do close to everything.
I am excited to live in a world where AI has "freed" humans from wage slavery, but our economic system is not ready to deal with that yet.
I'm skeptical. This will drastically change what it means to do a job in a way that has never happened before, but humans will find a way to deal with the fallout. We don't have a choice. Besides, if we were able to disrupt the very foundations of our economy for a minor virus, we can and will do the same to deal with this if required.
Either way this change has already arrived and we are starting to adapt our lives in response to it like we have many times in the past.
tldr: This change is significant but we'll manage.
I wouldn’t say the handling of COVID was smooth to say the least.
Yes we handled it, we are still paying the bill for that handling (inflation).
I think AI will have the disruption level of COVID, but there will not be an end in sight, 5%, 10, 20, 50% of people will lose jobs and even if they can refrain and handle it, it will take 5-10 years for those people to handle it. Can the countries have people on unemployment for that long ?
Productivity will skyrocket and with it the standard of living. Humans will always enjoy having other humans doing stuff for them.
Sure, it will be faster this time and there will be some growth pains.
It's not a matter of being ready, it's a matter of needing this. If you look at society's problems today, we're in a deadlock. I believe the benefits of AI can help alleviate a lot.
It will most likely widen, but who cares? What matters to me is the quality of my life, not others. If they're managing to get better than me while doing something useful to society, good for them.
What really matters is: the poor of tomorrow will laugh at the life of today's rich.
I mean, the poor won't have the Bezos' yatch, but they'll have access to some life amenities, health resources, etc, that Bezos can't even dream of having today.
Its alarmism to support government regulation to reinforce the moat when industry leaders say they intend to do it, but also that the danger of it being done is why competition with them must be restricted by the State (and why they can’t, despite being, or being a subsidiary of, a nonprofit founded on an openness mission, share any substantive information on their current models.)
But the concerns about AI taking over the world are valid and important; even if they sound silly at first, there is some very solid reasoning behind it. They’re big matrices, yes, but they’re Turing-complete which means they can theoretically do any computational task
See https://youtu.be/tcdVC4e6EV4 for a really interesting video on why a theoretical superintelligent AI would be dangerous, and when you factor in that these models could self-improve and approach that level of intelligence it gets worrying…
This comment basically implies I don't get it, but I will if I watch a Youtube video. I get it. ChatGPT isn't that. That's the point. You can have concerns about AGI. That's fine. But they have nothing to do with LLMs unless you are trying to play a shell game.
> They're big matrices and they are very cool tools!
Well, your mom is a etc
Edit: Since this is getting downvoted I'll be more explicit: The human brain may well be also just described as some simple sort of thing, but that doesn't mean humans are not dangerous, nor hypothetical humans with a brain ten times as large and a million times faster. The worry about AIs killing all humans soon is not naive just by sounding naive.
Sure, it's not naive just because it sounds naive. It's naive for other reasons (for one thing, we're really no closer to super-intelligent AIs than we were before the LLM craze began).
A lot of people would disagree with that. You can hardly deny that progress has sped up in the last few years, so I don't know why we shouldn't extrapolate this speed into the coming years.
"It is refreshing to hear opinions I already agree with. People with other opinions are unintelligent"
Is that what you were trying to convey? If not, I'm curious to know what you find refreshing about it and why those who disagree are wrapped in double quotes.
Well, it's to their benefit to portray their models as working alongside and enhancing humans, as opposed to replacing us. So it sounds a bit like marketing speak to me.
And it's to the benefit of many of those tech "thought leaders" to be alarmist since they don't have much of the AI pie
Note that this is business and economy academia, not real science. It's usually scientism at its peak: mathy explanations that use complex equations to describe poorly modeled phenomena and arrive at the previously decided on conclusion that the authors wanted to show. There's few other domains quite as smoke-and-mirrors as this (social sciences probably being the other, though they tend to at least steer clear of the complex math).
I have been following the market. Some of the most popular tech stocks have dropped back to pre-covid levels. During covid my inbox was blowing up from Meta, Amazon, Microsoft recruiters. This behavior clearly correlates with the stock evaluation. Literally everyone I know in tech was able to make good moves or get raises. Uncool companies had a hard time competing with SV giants.
I can cite the hiring frenzy, the stock evaluation, HUNDREDS of articles that disagree with his "SIMPLE" answer.
I work in this space (podcast ads) and the podcast ads tech maturity is years behind other media, like video, due to RSS. This lag is also reflected in that advertisers generally spend less on podcasting.
RSS makes it very difficult to measure the performance of an ad. Most players in this space wrap a publishers hosted podcast link with their own tracking link, and the main metric of engagement is a download. Downloads themselves are a pretty crude measurement and don’t hold up to something like minutes watched/listened.
Placing ads into the audio directly also just isn’t as performant as streaming it into the media like in video. Publishers and advertisers want more fine grained control over where the ad is placed, when it’s placed, and how it’s consumed by the listener. Podcast ads currently don’t deliver on that.
> Publishers and advertisers want more fine grained control over where the ad is placed, when it’s placed, and how it’s consumed by the listener.
There you go.
Ad tech is behind when it comes to podcasts? Who fucking cares. Not the audience. Podcast advertising is the closest thing to traditional broadcast ads for TV and radio—when ads were still inert instead of sentient. There's not a damn person on this side of the divide that sees the things you mention as being a significant downside. Not having the self-awareness to realize it is pure déformation professionnelle; this entire comment reads like a burglar lamenting the fact that people keeping their money in the bank instead of at home under their mattresses makes it harder to break in and take it.
The podcasters care? Because if the advertisers can't quantify ROI and audience, then they're less likely to spend their money with you, and your potential revenue is capped.
You, like the previous commenter, are stating the obvious. What's mystifying is the general tone suggesting that what either of you are saying is supposed to be insightful.
> The US-wide heat map aligns pretty well with agricultural and industrial pollution, not necessarily wildfires. Let's not turn every last thing into a climate change debate.
Does agricultural and industrial pollution not contribute to climate change?
Governments around the world impose various requirements on vehicle speeds, most notably ebikes being capped at 20mph for "safety reasons". Meanwhile private cars can travel at over 100mph for no apparent reason other than to not upset drivers at the expense of everyone else on the road.